Underwater acoustic communications
Related Articles
A deep learning framework based on structured space model for detecting small objects in complex underwater environments
Regular monitoring of marine life is essential for preserving the stability of marine ecosystems. However, underwater target detection presents several challenges, particularly in balancing accuracy with model efficiency and real-time performance. To address these issues, we propose an innovative approach that combines the Structured Space Model (SSM) with feature enhancement, specifically designed for small target detection in underwater environments. We developed a high-accuracy, lightweight detection model—UWNet. The results demonstrate that UWNet excels in detection accuracy, particularly in identifying difficult-to-detect organisms like starfish and scallops. Compared to other models, UWNet reduces the number of model parameters by 5% to 390%, substantially improving computational efficiency while maintaining top detection accuracy. Its lightweight design enhances the model’s applicability for deployment on underwater robots.
Acoustic impedance-based surface acoustic wave chip for gas leak detection and respiratory monitoring
Acoustic impedance enables many interesting acoustic applications. However, acoustic impedance for gas sensing is rare and difficult. Here we introduce a micro-nano surface acoustic wave (SAW) chip based on the acoustic impedance effect to achieve ultra-fast and wide-range gas sensing. We theoretically established the relationship between surface load acoustic impedance and SAW attenuation, and analyzed the influence of acoustic impedance on acoustic propagation loss under different gas/humidity media. Experimental measurements reveal that the differences in acoustic impedance generated by different gases trigger different acoustic attenuation, and can achieve wide-range (0–100 v/v%) gas monitoring, with ultra-fast response and recovery speeds reaching sub-second levels (t90 < 1 s, t10 < 0.5 s) and detection limit of ~1 v/v%. This capability can also be perfectly utilized for human respiratory monitoring, accurately reflecting respiratory status, frequency, and intensity. Consequently, the SAW chip based on the acoustic impedance effect provides a new solution for in-situ detection of gas leaks and precise monitoring of human respiration.
A unified acoustic-to-speech-to-language embedding space captures the neural basis of natural language processing in everyday conversations
This study introduces a unified computational framework connecting acoustic, speech and word-level linguistic structures to study the neural basis of everyday conversations in the human brain. We used electrocorticography to record neural signals across 100 h of speech production and comprehension as participants engaged in open-ended real-life conversations. We extracted low-level acoustic, mid-level speech and contextual word embeddings from a multimodal speech-to-text model (Whisper). We developed encoding models that linearly map these embeddings onto brain activity during speech production and comprehension. Remarkably, this model accurately predicts neural activity at each level of the language processing hierarchy across hours of new conversations not used in training the model. The internal processing hierarchy in the model is aligned with the cortical hierarchy for speech and language processing, where sensory and motor regions better align with the model’s speech embeddings, and higher-level language areas better align with the model’s language embeddings. The Whisper model captures the temporal sequence of language-to-speech encoding before word articulation (speech production) and speech-to-language encoding post articulation (speech comprehension). The embeddings learned by this model outperform symbolic models in capturing neural activity supporting natural speech and language. These findings support a paradigm shift towards unified computational models that capture the entire processing hierarchy for speech comprehension and production in real-world conversations.
High-performance magnetostatic wave resonators based on deep anisotropic etching of gadolinium gallium garnet substrates
Magnetostatic wave resonators based on yttrium iron garnet (YIG) are a promising technology platform for future communication filters. Such devices have demonstrated better quality factors than acoustic resonators in the 7 GHz range and above. However, the coupling coefficients of these resonators have been limited to less than 3%, primarily due to the restricted design space that is a result of microfabrication challenges related to the patterning of gadolinium gallium garnet (GGG), the substrate material used for growing single-crystal YIG. Here we report magnetostatic wave resonators created through the anisotropic etching of GGG substrates. Our approach, which is based on the YIG-on-GGG platform, uses a transducer with a hairclip-like structure. It is created by developing a microfabrication methodology that involves thinning and deep etching (up to 100 μm) of the GGG substrate. The resulting magnetostatic wave resonators exhibit a coupling of more than 8% in the 6–20 GHz frequency range.
Elastic trapping by acoustoelastically induced transparency
Elastic bound states in the continuum (BICs) have recently attracted significant interests due to their exceptionally high-Q-factor, which enables the confined mode to be completely decoupled from spectrally coexisting radiative channels. We report on the emergence of a state that induces a slow vibration phenomenon, which exhibits a multiphysics analogy to the notion of slow light observed in electromagnetically induced transparency (EIT). Such a state can be achieved through the interaction of acoustoelastic coupling. Our proposed design involves a composite with two acoustic cavities encased in an elastic bar, making quasi-BICs feasible with high spatial efficiency in a localized area while allowing for the tunability of the Purcell factor by around six orders of magnitude. The observation of quasi-BICs with acoustoelastically induced transparency (AEIT) lineshapes, which are manifested by the coupling of two disparate physics domains, will expand the BIC family and enable applications in areas such as lasing, sensing, screening, and energy storage platforms where ultrahigh-Q-factor modes and radiative channels coexist.
Responses